Generation and Description of Neuronal Morphology & Connectivity This continuing project is directed at describing neuroanatomical structure in a compact yet sufficiently complete fashion to allow the implementation of biologically plausible and quantitatively accurate computer simulations. Neuronal morphology plays a fundamental role in physiological and pathological brain function by integrating complex patterns of synaptic inputs, transmitting trains of spiking output, and subserving network connectivity. During the previous funding periods (under Generation and Description of Dendritic Morphology), informatics tools were successfully designed and deployed to reproduce the three-dimensional shape of dendritic trees in the same format used to represent experimentally reconstructed neurons. Digital arbors were also combined with computational models of membrane biophysics to investigate the cellular structure-activity relationship. The goal of this application is to expand these software resources and research approach from dendrites to all aspects of neuronal structure, including full axonal arborizations and synaptic connectivity. The general strategy is to resample in stochastic models the experimentally measured statistical distributions, and to compare the resulting simulations directly to the original data. Such comprehensive and parsimonious characterization constitutes an effective way to compress, store, exchange, and amplify extremely complex neuroanatomical information. The project has three logically related, but technically independent specific aims. The first aim is to enhance the power and usability of computational neuroanatomy tools for the analysis and synthesis of neuronal morphology, and to integrate them with leading bioinformatics algorithms enabling large scale knowledge mining of massive data sets. In the second aim, digital reconstruction, quantitative morphometry, and compartmental modeling of branch growth and spike propagation are applied to two distinct classes of axonal arbors, namely hippocampal CA3 interneurons and olivo-cerebellar climbing fibers. The third aim, extending to circuitry, develops a relational database of cellular-level connectivity in the rodent hippocampus. In this framework, population statistics for each neuronal class are stochastically resampled to quantify the network structure-activity relationship. The neurobiological and technological components of this project are deeply intertwined and span a variety of scientific approaches, including microscopic imaging, computational simulations, statistical analysis and data mining. The robust development and open source distribution of the underlying neuroinformatics infrastructure for data handling and integration will continue to benefit the wider neuroscience community.